The present invention relates to integrated circuits and fabrication techniques therefor, and more particularly, to memory devices and methods of fabrication therefor.
Factors, such as an ongoing desire for increased circuit integration and the development of new device configurations (e.g., new memory cell types) that are scaleable to extremely small dimensions, have led to an increasing need for techniques for fabricating very small features in integrated circuit devices. Lower bounds on feature size often arise from constraints of photolithography, i.e., the resolution to which layers can be patterned and properly aligned may constrain the size of features that can be fabricated. Techniques, such as the use of sidewall spacers to reduce patterned dimensions of objects like holes in material layers and the use of creative etching techniques, have been developed to lessen some of these constraints, but other barriers to reliably and repeatably forming small structures still remain.
Damascene processes are commonly used in integrated circuit processing to form features such as contacts and wiring patterns. For example, in a typical conventional damascene process, a silicon dioxide layer is formed on a microelectronic substrate. A groove (for wiring) and/or an opening to an underlying region (for a contact) is formed in the dielectric layer. A conductive layer (e.g., a metal containing layer) is then deposited on the dielectric layer, filling the groove and/or opening. Chemical mechanical polishing (CMP) may then be used to remove portions of the conductive layer disposed on the dielectric layer, thus leaving a wiring pattern in the groove and/or a contact plug in the opening.
Such techniques may be used, for example, in fabricating a lower electrode contact (or “small contact”) that provides a high current density path for heating a phase-changeable material (e.g., chalcogenide) region in a phase-change memory device. In a typical fabrication process for such a cell, a dielectric layer is formed over a conductive plug or pad that is electrically coupled to a source/drain region of an access transistor formed on a semiconductor substrate, and a small contact hole is made in the dielectric layer to expose an upper surface of the plug or pad. A metal-containing material is then deposited on the dielectric layer and in the small contact hole. Excess material disposed on the dielectric layer is then removed using CMP to leave a small contact plug in the contact hole. A phase-changeable material region is then formed on the surface of the dielectric layer and the small contact plug, and an upper electrode is formed on the phase-changeable material region. Examples of techniques for forming contacts for phase-change memory devices are described in U.S. Pat. No. 6,117,720 and U.S. Pat. No. 6,147,395.
Conventional processes may have characteristics that can limit the ability to reliably and repeatably make small contacts or other small structures. In particular, in many applications, it may be desirable to remove a metal or other conductive layer as close as possible to the top of a surrounding dielectric layer or region. For example, in forming small contact plugs for phase-change memory cells along the lines described above, it is generally desirable to remove the metal layer down to a shoulder of the opening in the dielectric layer so that the surface area of the individual contact plugs is made as small as possible while maintaining the planarity of the substrate surface and uniformity among the contact plugs. However, using a conventional process as described above can result in less than desirable results due to flaring at the mouths of the contact holes and/or dishing, overerosion, edge over-erosion, and other surface non-uniformity arising from the CMP. Such effects may be exacerbated by variation in pattern density across the surface of the wafer.